John Thurlow Wilson.

Effects of individual differences on learning from written materials: control of inspection behavior by test-like events. online

. (page 5 of 8)
Online LibraryJohn Thurlow WilsonEffects of individual differences on learning from written materials: control of inspection behavior by test-like events. → online text (page 5 of 8)
Font size
QR-code for this ebook

performance relationships.

Instructional Treatment Main Effects

Acquisition of relevant information . — The first
hypothesis tested was:

1. Subjects receiving treatments with adjunct ques-
tions inserted into the written passage will
exhibit greater acquisition of relevant information
than the control group.
As previously defined, information is identified as
relevant when it is the intended response to an inserted
question. Therefore, when a subject can respond correctly
to the relevant post-test questions, which are identical
with the inserted question, his performance is evidence for
the acquisition of relevant information. Support for this
hypothesis is dependent upon significant between group
differences in scores on relevant post-test questions. The
appropriate statistical tests of this hypothesis are there-
fore comparisons of performances on relevant text or


diagram questions by pairs of treatment conditions follow-
ing an overall significant F ratio. The data in Table 14
strongly supported this hypothesis.

In general, when subjects answered questions within
the written passage, they also performed better on those
same questions on the post-test than subjects who did not
receive the questions within the written passage. Hence,
subjects who received inserted diagram questions also
performed better on the relevant diagram questions on the
post-test. The same was true of subjects receiving inserted
text questions in that their performance was better than
other subjects on relevant text questions on the post-test.
This finding is consistent with earlier studies (Rothkopf,
1966} Rothkopf and Bisbicos, 1967? Frase, 1970) where it
was shown that the insertion of questions in written instruc-
tional material increases the amount learned from the text
when they occur after the written passage to which they

The facilitative effects of questions appearing after
the written passage upon the acquisition of relevant informa-
tion is often identified as a direct instructional effect of
questions. This effect recognizes that the content load of
the question itself provides additional opportunity for the
acquisition of that information in subsequent questions with
similar content loads. The repetition and practice learners
perform when responding to content- loaded questions would
suggest that a relationship should exist between answering


the question correctly and the performance on the relevant
post-test questions. This expectation was confirmed by
regression analysis where positive relationships were
disclosed between the number of inserted questions answered
and the performance on the corresponding relevant questions.
This relationship occurred for both inserted diagram and
inserted text questions treatment conditions (Figure 4 and


In summary, both the significant between-group differ-
ences aad the positive relationships between performances
on inserted questions and relevant post-test questions
support the hypothesis that inserted questions facilitate
acquisition of the relevant information from both diagrams
and text. In addition, the mechanism involved in both
cases can be identified as the direct instructional effects
of the content load of the inserted questions on inspection
behavior and subsequent acquisition of similar content in
passages that follow.

Acquisition of incidental information .— The second
and third hypotheses tested were:

2. Subjects receiving treatments with adjunct

questions based upon diagrammatic information
inserted into the written material will exhibit
significantly greater acquisition of incidental
information from diagrams than groups that did
not receive diagrammatic adjunct questions.


3. Subjects receiving treatment with adjunct ques-
tions based upon textual information inserted
into the written passage will exhibit significantly
greater acquisition of incidental information from
text than groups that did not receive textual
adjunct questions.
Information is considered to be incidental when it is
not an intended response to the questions inserted into the
written passages. Acquisition of incidental information
is measured by performance on post-test questions that
are not the same as those inserted into the written passage.
While the inserted questions also appearing on the post-test
are called relevant post-test questions, these are called
incidental post-test questions. Support for these hypotheses
is dependent upon significant between-group differences in
scores on incidental post-test questions. The appropriate
statistical tests of these hypotheses are therefore compari-
sons of performances on incidental text or diagram questions
by pairs of treatment conditions following an overall
significant F ratio. The data in Table 14 were collected
to test these hypotheses. However, the non-significant F
ratios for between-group differences on Incidental post-test
questions for both diagram and text information fail to
support hypotheses 2 and 3.

Evidence for incidental learning is measured on post-
test questions that are not the same as those inserted into
the text. Therefore, two additional performances should be


considered; first, performances on relevant text questions
when the treatment condition was inserted diagram questions,
and second, performances on relevant diagram questions when
the treatment condition was inserted text questions. The
comparisons of these performances reported in Table 15 and
16 disclose that treatments with inserted text questions did
not perform significantly better on the relevant diagram
questions than treatment conditions with no inserted
questions. The lack of significance was also found between
treatments with inserted diagram questions and no inserted
questions relative to their performance on the relevant
text questions. This additional evidence also fails to
support hypotheses 2 and 3.

The effects of inserted questions in written passages
provides a means to contrast acquisition of relevant and
incidental learning. While the analysis of variance and
subsequent comparisons proved to be insignificant, regres-
sion analysis between the number of inserted questions and
the performance on another set of relevant post-test
questions was generally positive. Here, treatments with
inserted text questions displayed a positive relationship
to performance on relevant diagram questions but regression
slopes were somewhat less steep and not as high as treatments
with inserted diagram questions (Figure 4) . A similar
relationship (Figure 7) occurs between treatment conditions
with inserted diagram questions and their performance
relative to relevant text questions. These observations


are consistent with the observation that an additional
multiplier is involved when considering the relative
facilitative effects of inserted questions upon relevant
versus incidental learning.* This multiplier refers to the
fact that the relevant information, as measured by the
relevant post-test questions, represents fairly limited
sized population of content while the incidental informa-
tion, represents a much larger sized universe of content.
Subjects in the treatment groups with the relevant inserted
questions are cued to the Information necessary for acquisi-
tion in order to produce better post-test performances.
Therefore, they can attend to less information than subjects
in other treatments would have to consider in order to
achieve the same post-test performance. It is therefore
hard to tell how powerful the influence of inserted ques-
tions really is on these two performances, even when the
post-test questions are the same.

In summary, the evidence collected does not generally
support hypotheses 2 and 3. However, inserted questions
may have actually exerted a great influence on the acquisi-
tion of incidental information. Partial explanation of
this lies in the fact that the incidental post-test questions
have sampled only a very small portion of the incidental
content as compared to the 100 percent sampling of the
relevant content.

*E. Z. Rothkopf, personal communication.


Highlighting effects .— The fourth hypothesis tested


4. There will be a significant influence on performance
for subjects who receive treatments in which they
highlight information in the written materials.
Support for this hypothesis is dependent upon signifi-
cant between group differences in post-test scores for groups
that highlighted and those that did not. The appropriate
statistical test of this hypothesis is a significant F
ratio. Data in Table 14 indicated that subjects who did
not highlight diagrams performed significantly better on
the relevant diagram questions than those who did highlight
diagrams. No significant differences were found between
those who highlighted text and those who did not, even
though the means for highlighting text groups were generally
lower than the non-highlighting text groups,

Turther analysis of highlighting activity by applying
a repeated measures analysis using the amount of information
highlighted as the repeated measure across written passages
revealed that the change in the percent of information
highlighted from the base rate for passage one increased
for diagrams but decreased for texts. Another analysis
using regression techniques comparing time spent in the
treatment to performance on relevant post-test questions
revealed that highlighting activity and performance were
confounded with the time spent in the treatment. Generally,
when subjects in treatment conditions that included


highlighting activity spent little time in the treatment,
their performances were generally poorer on the post-test
questions than subjects who spent more time. However,
when they spent less time without highlighting, their
performances improved (Figure 3) . Since highlighting
activity for diagrams increased across the written passages,
the amount of time required to do the highlighting would
also increase. The chance exists that the activity of
highlighting may occur without attending to detail and in
some cases the subject's attention may have been placed on
the activity of highlighting itself.

When highlighting and no highlighting were compared
using regression analysis for treatment conditions without
questions, an interesting, potential disordinal interaction
occurred (Figure 3) between treatment time and performance
on relevant diagram questions. When a short time was spent
studying written materials (treatment time) , subjects
benefited most by not spending the extra effort or time
trying to highlight important information. However, high-
lighting provided a chance to facilitate acquisition when
study time in the treatment exceeded about twenty minutes.
While highlighting may hinder the subject's inspection
activity during a short study time, highlighting seemed to
facilitate performance during longer time periods. This
facilitative effect may be due to the influence of high-
lighting activity to maintain the focus of attention during
long periods of time when the tendency would be to shift the
focus of attention to distractions.


In sumnary, highlighting activity seems to inhibit
performance on relevant post-test questions, according to
differences between group means. However, when highlighting
occurred without any inserted questions, a positive relation-
ship occurred between time and relevant post-test performance,
indicating that highlighting potentially may exert an influence
upon inspection behavior in a similar manner as inserted ques-
tions. When combined with inserted questions, the effects of
highlighting are not always positive and in some cases in-
hibit performance, as reflected by differences between mean
performances. While the beneficial effects of highlighting,
when they occur, have been interpreted as a means to focus
attention to important specifics, this type of treatment was
found to interfere with the achievement of high performances.
This finding is consistent with studies where a high amount
of repetition or attention to detail fails to produce the
higher performance (Rothkopf and Coke, 1966; Rothkopf, 1968).
Evidentally, in the practice of highlighting activity, the
learner may read the passage and then go back and highlight
(often immediately) what he considered to be important. It
may well be that this unmeasured delay in highlighting, when
it occurs, is an important variable confounded with
Aptitude X Treatment Interactions

The fifth hypothesis tested was:

5. There will be a differential relationship between

criterion performances and aptitudes of subjects,

relative to the treatment received.


This hypothesis implies disordinal interactions
between aptitude and treatment. More specifically, the
regression line relating aptitude to criterion scores
under one treatment intersects the regression line for the
alternative treatment. The statistical tests of this
hypothesis are significant F tests for heterogeneity of
regression (Cronbach and Snow, 1969) .

The evidence with respect to hypothesis 5 consists
primarily of interactions between treatment time and post-
test performance and one aptitude factor test interacting
with post-test performance. While treatment time -may seem
to be an unlikely aptitude, aptitude has been defined as
any characteristic of the learner which facilitates or
interferes with his learning from some designated instruc-
tional method (Cronbach and Snow, 1969) .

Treatment time . — Analysis of treatment time with
respect to post-test performance revealed that treatments
with text questions without highlighting and no inserted
questions with highlighting displayed a consistently
positively relationship with performance on both relevant
and incidental post-test questions. Hence, subjects who
spent more time in the treatment generally scored higher
when assigned to these treatment conditions. Of special
interest are treatments of no inserted questions without
highlighting, inserted text questions without highlighting,
and inserted diagram questions without highlighting which
demonstrated consistently negative relationships between


treatment time and post-test performance on relevant and
incidental information as well as on an overall post-test
performance. When subjects spent more time in these treat-
ments, their performances attenuated.

Taken as a whole , the patterns of treatment time and
performance relationships suggests that the effects of
questions inserted into the written passages or highlighting
can facilitate acquisition. However, when both are present,
a heavy burden is placed upon the physical doing within the
treatment materials producing in some cases a poorer per-
formance suggesting that the treatment has interfered with
effective processing of the material or attenuated perform-
ance through boredom or fatigue. A similar effect was
observed for treatments with only inserted diagram questions
and no inserted questions. Evidentally, these conditions
provide little influence and, as a result the learner adapts
less productive behaviors, at least in terms of time.

In summary, how a learner spends his time acquiring
information from instructional material is of basic impor-
tance when it can be shown that time and performance are
concomitant variable within instructional conditions.

Aptitude factors . —The initial aptitude factors for
this study were selected on the basis of an analysis of
learning from diagrammatic and textual mode of instructional
material, identifying the tasks and processes corresponding
to a model proposed by Melton (1967) for the investigation
of learning and individual differences. A summary of


individual differences, identified by the aptitude factor
tests utilized, is presented below (sununarized from page 26)

Sj^ >• J^i^Si^

Hidden Patterns Object-Number _

Number Comparison Auditory Letter Span


Organization of Aptitude Factor Tests
in a Multi-process Model of Learning

Of the five aptitude factors studied, only the object-
rtumber factor test was found to significantly interact with
treatment conditions, relative to performance on the
relevant text questions. An analysis of this interaction
with the relevant text questions (Figure 5) indicates that
subjects who scored high on the object-number test also
benefited most from treatments with text questions with or
without highlighting activity and least from diagram ques-
tions without highlighting-

The object-number test of associative memory presents
explicit pairs of objects and numbers that the subjects are
asked to learn. They are told, in taking the test, that
they will be allowed to study the list of object-number
pairs and then will be asked to write the number for the
same objects, but only arranged in a different order. Hence,
the nature of the associations to be formed is quite clear
for this particular test as well as for other tests of
associative memory (French, Ekstrom, and Price, 1963).
However, in learning from written materials, subjects must


identify the meaningful associative pairs within or between
sentences and then commit them to memory. In the present
experiment, questions could conceivably be formed from
nearly every sentence. The cues provided by text questions
identify important associated information and permit
subjects to capitalize on associative memory. In other
words, the insertion of questions serve to clarify the
nature of the associations to be formed and thus allow
associative memory to be more effectively utilized. This
finding is consistent with other research (Koran, M. L. and
Koran, J- J., 1972) where treatment conditions with high
frequency of inserted questions displayed a positive rela-
tionship between incidental post-test questions and the
scores on the first-last names test of associative memory.
The fact that individual differences have more often been
found for incidental learning than relevant learning has
been presumed to be caused by the higher motivation to
learn relevant information, thus obscuring differences in
learner habits and sets (Plunderleith and Postman, 1957).
However, most other research that has found these differences
for incidental learning utilized college aged subjects,
rather than the younger high school aged subjects used in
this study. It is possible that in the latter case
subjects are less motivated to participate, especially when
the written materials are not part of their regular school
work. It is also possible that the high level of the ques-
tion difficulty has also limited the effectiveness of
inserted questions.



The relationship between associative memory and
processing of textual material was predicted in the analysis
in Chapter I using Melton's Model (1967). Another relation-
ship was also predicted between perceptual speed and learning
from diagrams. Specifically, diagrams can be described as
a presentation of specifics within an organizational schema
of associations, namely the proximity and sequence of the
specifics within the diagram. Here, stimulus differentia-
tion is required in order to assess associations between
symbols. Acquisition becomes dependent upon the subjects'
perceptual speed and accuracy in scanning these specifics.
The number comparison test was identified as a suitable
measure of perceptual speed and accuracy. In this test,
subjects are presented with 50 pairs of numbers (usually
longer than six digits each) . The subject was to identify
those pairs that are not the identical pairs in a very
short period of time. Here also both speed and accuracy of
perception are important.

In Table 31 , an interaction of borderline significance
is reported between number comparison factor -test and treat-
ment conditions, in relation to performances on relative
diagram questions. An analysis of this interaction indicates
that performances in treatment conditions with diagram ques-
tions with highlighting generally performed better but without
a strong relationship to perceptual speed and accuracy. The
same was found for all other treatments with highlighting.
Evidentally, highlighting activity as well as diagram


questions minimize the demands upon perceptual speed and
accuracy and permit subjects to capitalize on other
aptitudes in the processing of the diagrammatic information.
VJhen these prompts are not present in the treatment condi-
tions, such as inserted text questions without highlighting
or no questions with no highlighting, performances place
a -higher burden on perceptual speed and accuracy, as
evidenced by steeper slopes. These latter treatments
require subjects to find and accurately identify important
information within the diagram without support of the
treatment condition.

In summary, highlighting in general inhibited the
acquisition of relevant diagram information. However, it
was found to benefit certain low ability learners, perhaps
by reducing certain processing requirements as well as
enabling them to capitalize upon others. In this sense,
the treatment condition provided compensation for defi-
cient attentional and perhaps discriminational skills.

I ndividual Differences in Learning Re levant
■ and Incidental Information

This research has attempted to further examine the
relationships between individual differences and inspection
behavior relative to learning from different modes of
instructional content. In order to search out these relation-
ships, treatments must be designed that capitalize upon the
individual learner's habits and learning set. Treatments
that strongly influence these habits and learning set usually


fail to expose learner characteristics and performance
relationships. Therefore, the fact that the treatments
in this study displayed no major significant advantage for
one over another indicates that they may be suited for the
investigation of aptitude x treatment interaction.

Interesting direct instructional effects disclosed
that inserted questions with their intended responses
selected from diagrams were an effective means to influence
acquisition of the intended , relevant information from the
diagram. An additional phenomena was observed. Answering
most of the inserted text questions also facilitated acquisi-
tion of diagrammatic information. However, answering most
of the inserted diagram questions did not necessarily
facilitate acquisition of relevant text information, unless
the task also included the performance of highlighting
activity. This observation leads to the possibility that
the processing activities are performed differentially.

Generally, processing activity refers to the manipula-
tions performed by the learner in transforming the physical
stimulus to an encoded stimulus. When he inspects and
attends to some material, it is assumed that this action
influences the nature of the encoded stimulus. While
inspection and other processing activities are considered
to be sensitive to external influence, the effects of this
external influences seems to occur differentially, relative
to the individual differences in aptitude and performance.
In addition, the effects of different modes of instructional
content seem to increase these differential effects. Hence,


learners may be performing one type of processing activity
with diagrams and another for texts.

How inspection behavior is influenced and shaped by
external influences, such as inserted questions or high-
lighting, has remained without general consensus (Watts and
Anderson, 1971; McGaw and Grotelueschen, 1972). Considering
the seemingly conflicting roles inserted questions and
highlighting seem to exhibit, inspection behavior may or
may not be influenced in some of these cases, and therefore
may be dependent upon the mode of information, the difficulty
of the material and questions, and the relevant learner
characteristics. It is very conceivable that inspection
behavior alone is an inadequate explanation. Watts and
Anderson (1971) have suggested that the review of informa-
tion a learner performs when answering a question also
influences his performance of the acquisition of information.
In this study, review in answering inserted text questions
has possibly helped in the acquisition of relevant diagram
Information. However, the review when answering inserted
diagram questions did not demonstrate the same effect on
acquisition of relevant text information. Here only when

1 2 3 5 7 8

Online LibraryJohn Thurlow WilsonEffects of individual differences on learning from written materials: control of inspection behavior by test-like events. → online text (page 5 of 8)